主管:中国科学院
主办:中国优选法统筹法与经济数学研究会
   中国科学院科技战略咨询研究院

中国管理科学 ›› 2023, Vol. 31 ›› Issue (9): 22-34.doi: 10.16381/j.cnki.issn1003-207x.2022.0998cstr: 32146.14.j.cnki.issn1003-207x.2022.0998

• • 上一篇    下一篇

考虑基金网络地位的多阶段效率及其前瞻性预测能力

陆帅1,2, 李守伟1,2(), 何建敏1   

  1. 1.东南大学经济管理学院,江苏 南京 211189
    2.东南大学系统与信息工程研究发展中心,江苏 南京 211189
  • 收稿日期:2022-05-07 修回日期:2022-08-30 出版日期:2023-09-15 发布日期:2023-09-19
  • 通讯作者: 李守伟 E-mail:lishouwei@seu.edu.cn

Multistage Efficiency Considering Network Status of Funds and its Predictive Ability

Shuai LU1,2, Shou-wei LI1,2(), Jian-min HE1   

  1. 1.School of Economics and Management,Southeast University,Nanjing 211189,China
    2.Research and Development Center for System and Information Engineering,Southeast University,Nanjing 211189,China
  • Received:2022-05-07 Revised:2022-08-30 Online:2023-09-15 Published:2023-09-19
  • Contact: Shou-wei LI E-mail:lishouwei@seu.edu.cn

摘要:

近年来,常被用于测度商业银行运营效率的两阶段DEA方法逐渐出现在开放式基金绩效评价研究中。然而,相较于银行绩效,开放式基金绩效具有显著的网络地位效应,即在社会网络中具有更高地位的开放式基金可能利用其影响力传播噪音,达到掩盖负面消息的目的。已有研究大多在基金绩效评价时忽略了基金网络地位因素,从而造成评价的偏误。对此,本文利用复杂网络方法构建了基金的共同持股网络,测算开放式基金的混合网络中心性以刻画其网络地位,并将基金网络地位加入两阶段DEA模型形成新的多阶段投入产出框架,从而得到考虑基金网络地位的多阶段DEA模型。基于该模型,本文测度了2015-2020年间中国2600支开放式基金的考虑基金网络地位的多阶段效率NSE,并检验NSE对于基金未来绩效的前瞻性预测能力。结果表明,NSE对于未来超额收益率和总资产回报率具有显著正向影响,而对于最大涨幅无显著影响;NSE对于未来系统性风险和年化波动率具有显著正向影响,而对于最大回撤无显著影响。采用倾向得分匹配法缓解样本自选择问题后结论未变。同时,本文还发现NSE对于基金未来绩效的预测能力在不同现金流与投资集中度的基金中存在显著异质性影响。稳健性测试支持了本文主要结论。

关键词: 基金网络地位, 网络中心性, DEA, 预测能力, 多阶段效率测度

Abstract:

The traditional two-stage DEA method, which measures the operating efficiency of commercial banks, has gradually been used in related research on the performance evaluation of funds. However, compared with bank performance, mutual funds have a significant network status effect, that is, funds with higher status in social networks may perform better due to signal transmission effect. The key issue is that most of the research has ignored the fund network status factors, leading to errors in the evaluation. To fill this gap, complex network model is employed to build a common-holding network of funds and calculate the mixed network centrality of funds to portray their network status. Specifically, the mixed network centrality is composed of the average of degree centrality, betweenness centrality and eccentricity. Based on the mixed network centrality, the fund network status is introduced into the two-stage DEA model to construct a new multi-stage input-output framework considering the fund network status.Based on this model, the multi-stage efficiency considering the network status (NSE) of 2600 mutual funds in China from 2015-2020 is measured and tested the predictability of NSE for future fund performance. The results show that NSE has a significant positive impact on future abnormal return and ROA; NSE has a significant positive impact on future systemic risks and annualized volatility, while it has an insignificant effect on the maximum retracement. The conclusions keep unchanged after using the PSM method to relieve the problem of the sample selection. It is also found that the predictability of NSE for future performance presents significant heterogeneity in funds with different cash flow and investment concentrations. Last, robustness checks support the main conclusions. The financial data of mutual funds is from the Wind database in China. A new approach is provided to measure multi-stage efficiency by considering the fund network status that are significant for evaluating and predicting fund performance. It also calls for more empirical or theoretical studies to detect social connections of mutual funds and evaluate their real efficiency performance.

Key words: fund network status, network centrality, DEA method, predictive power, multistage efficiency measure

中图分类号: